Abstract/Summary

Variability in the Arctic stratospheric polar vortex can significantly influence the behaviour and predictability of wintertime tropospheric weather conditions on subseasonal timescales (lead-times between 2 weeks and 2 months). This thesis improves our understanding of tropospheric processes leading to subseasonal stratospheric variability and how to characterise the tropospheric impact of stratospheric variability in ways useful for subseasonal prediction. First, the predictability onset of the February 2018 sudden stratospheric warming is linked to the occurrence of synoptic-scale anticyclonic wave breaking in the northeast Atlantic. This event is diagnosed through the Scandinavia-Greenland dipole pattern in mean sea-level pressure. Analysis of previous cases shows the Scandinavia-Greenland pattern is associated with significantly enhanced vertically propagating wave activity and a subsequently weakened polar vortex. The representation of the Scandinavia-Greenland pattern in hindcasts from the S2S Prediction Project database is then analysed. Substantial biases are found in the representation of the pattern in the troposphere and its relationship with a weakened polar vortex. These biases likely contribute to limiting subseasonal stratospheric forecast skill and highlight the need to improve the representation and predictability of the Scandinavia-Greenland pattern for improved polar vortex forecasts. Next, the downward influence of the stratosphere on the troposphere over North America is characterised using four weather regimes, defined through clustering analysis. Significant observed differences in regime behaviour are found for three of the four regimes depending on the strength of the lower-stratospheric vortex. By considering the empirical orthogonal functions which define the regimes, a linear theory is developed to explain how changes in the stratospheric vortex strength can lead to regime transitions. The theory is supported by the results from a set of stratospheric relaxation model experiments. These results provide a framework to explain how uncertainty or improvements in the subseasonal stratospheric forecast can translate to changes in large-scale tropospheric forecasts over North America.